1 from __future__ import unicode_literals
2
3 import itertools
4 import json
5 import math
6 import operator
7 import re
8
9 from .utils import (
10 error_to_compat_str,
11 ExtractorError,
12 js_to_json,
13 remove_quotes,
14 unified_timestamp,
15 )
16 from .compat import (
17 compat_collections_chain_map as ChainMap,
18 compat_itertools_zip_longest as zip_longest,
19 compat_str,
20 )
21
22 _NAME_RE = r'[a-zA-Z_$][\w$]*'
23
24 _UNDEFINED = object()
25
26
27 def _js_bit_op(op):
28
29 def wrapped(a, b):
30 def zeroise(x):
31 return 0 if x in (None, _UNDEFINED) else x
32 return op(zeroise(a), zeroise(b))
33
34 return wrapped
35
36
37 def _js_arith_op(op):
38
39 def wrapped(a, b):
40 if _UNDEFINED in (a, b):
41 return float('nan')
42 return op(a or 0, b or 0)
43
44 return wrapped
45
46
47 def _js_div(a, b):
48 if _UNDEFINED in (a, b) or not (a and b):
49 return float('nan')
50 return float('inf') if not b else operator.truediv(a or 0, b)
51
52
53 def _js_mod(a, b):
54 if _UNDEFINED in (a, b) or not b:
55 return float('nan')
56 return (a or 0) % b
57
58
59 def _js_exp(a, b):
60 if not b:
61 # even 0 ** 0 !!
62 return 1
63 if _UNDEFINED in (a, b):
64 return float('nan')
65 return (a or 0) ** b
66
67
68 def _js_eq_op(op):
69
70 def wrapped(a, b):
71 if set((a, b)) <= set((None, _UNDEFINED)):
72 return op(a, a)
73 return op(a, b)
74
75 return wrapped
76
77
78 def _js_comp_op(op):
79
80 def wrapped(a, b):
81 if _UNDEFINED in (a, b):
82 return False
83 return op(a or 0, b or 0)
84
85 return wrapped
86
87
88 # (op, definition) in order of binding priority, tightest first
89 # avoid dict to maintain order
90 # definition None => Defined in JSInterpreter._operator
91 _DOT_OPERATORS = (
92 ('.', None),
93 # TODO: ('?.', None),
94 )
95
96 _OPERATORS = (
97 ('>>', _js_bit_op(operator.rshift)),
98 ('<<', _js_bit_op(operator.lshift)),
99 ('+', _js_arith_op(operator.add)),
100 ('-', _js_arith_op(operator.sub)),
101 ('*', _js_arith_op(operator.mul)),
102 ('/', _js_div),
103 ('%', _js_mod),
104 ('**', _js_exp),
105 )
106
107 _COMP_OPERATORS = (
108 ('===', operator.is_),
109 ('==', _js_eq_op(operator.eq)),
110 ('!==', operator.is_not),
111 ('!=', _js_eq_op(operator.ne)),
112 ('<=', _js_comp_op(operator.le)),
113 ('>=', _js_comp_op(operator.ge)),
114 ('<', _js_comp_op(operator.lt)),
115 ('>', _js_comp_op(operator.gt)),
116 )
117
118 _LOG_OPERATORS = (
119 ('|', _js_bit_op(operator.or_)),
120 ('^', _js_bit_op(operator.xor)),
121 ('&', _js_bit_op(operator.and_)),
122 )
123
124 _SC_OPERATORS = (
125 ('?', None),
126 ('??', None),
127 ('||', None),
128 ('&&', None),
129 )
130
131 _OPERATOR_RE = '|'.join(map(lambda x: re.escape(x[0]), _OPERATORS + _LOG_OPERATORS))
132
133 _MATCHING_PARENS = dict(zip(*zip('()', '{}', '[]')))
134 _QUOTES = '\'"/'
135
136
137 def _ternary(cndn, if_true=True, if_false=False):
138 """Simulate JS's ternary operator (cndn?if_true:if_false)"""
139 if cndn in (False, None, 0, '', _UNDEFINED):
140 return if_false
141 try:
142 if math.isnan(cndn): # NB: NaN cannot be checked by membership
143 return if_false
144 except TypeError:
145 pass
146 return if_true
147
148
149 class JS_Break(ExtractorError):
150 def __init__(self):
151 ExtractorError.__init__(self, 'Invalid break')
152
153
154 class JS_Continue(ExtractorError):
155 def __init__(self):
156 ExtractorError.__init__(self, 'Invalid continue')
157
158
159 class JS_Throw(ExtractorError):
160 def __init__(self, e):
161 self.error = e
162 ExtractorError.__init__(self, 'Uncaught exception ' + error_to_compat_str(e))
163
164
165 class LocalNameSpace(ChainMap):
166 def __getitem__(self, key):
167 try:
168 return super(LocalNameSpace, self).__getitem__(key)
169 except KeyError:
170 return _UNDEFINED
171
172 def __setitem__(self, key, value):
173 for scope in self.maps:
174 if key in scope:
175 scope[key] = value
176 return
177 self.maps[0][key] = value
178
179 def __delitem__(self, key):
180 raise NotImplementedError('Deleting is not supported')
181
182 # except
183 def pop(self, key, *args):
184 try:
185 off = self.__getitem__(key)
186 super(LocalNameSpace, self).__delitem__(key)
187 return off
188 except KeyError:
189 if len(args) > 0:
190 return args[0]
191 raise
192
193 def __contains__(self, key):
194 try:
195 super(LocalNameSpace, self).__getitem__(key)
196 return True
197 except KeyError:
198 return False
199
200 def __repr__(self):
201 return 'LocalNameSpace%s' % (self.maps, )
202
203
204 class JSInterpreter(object):
205 __named_object_counter = 0
206
207 undefined = _UNDEFINED
208
209 RE_FLAGS = {
210 # special knowledge: Python's re flags are bitmask values, current max 128
211 # invent new bitmask values well above that for literal parsing
212 # TODO: new pattern class to execute matches with these flags
213 'd': 1024, # Generate indices for substring matches
214 'g': 2048, # Global search
215 'i': re.I, # Case-insensitive search
216 'm': re.M, # Multi-line search
217 's': re.S, # Allows . to match newline characters
218 'u': re.U, # Treat a pattern as a sequence of unicode code points
219 'y': 4096, # Perform a "sticky" search that matches starting at the current position in the target string
220 }
221
222 _EXC_NAME = '__youtube_dl_exception__'
223 _OBJ_NAME = '__youtube_dl_jsinterp_obj'
224
225 OP_CHARS = None
226
227 def __init__(self, code, objects=None):
228 self.code, self._functions = code, {}
229 self._objects = {} if objects is None else objects
230 if type(self).OP_CHARS is None:
231 type(self).OP_CHARS = self.OP_CHARS = self.__op_chars()
232
233 class Exception(ExtractorError):
234 def __init__(self, msg, *args, **kwargs):
235 expr = kwargs.pop('expr', None)
236 if expr is not None:
237 msg = '{0} in: {1!r}'.format(msg.rstrip(), expr[:100])
238 super(JSInterpreter.Exception, self).__init__(msg, *args, **kwargs)
239
240 @classmethod
241 def __op_chars(cls):
242 op_chars = set(';,')
243 for op in cls._all_operators():
244 for c in op[0]:
245 op_chars.add(c)
246 return op_chars
247
248 def _named_object(self, namespace, obj):
249 self.__named_object_counter += 1
250 name = '%s%d' % (self._OBJ_NAME, self.__named_object_counter)
251 namespace[name] = obj
252 return name
253
254 @classmethod
255 def _regex_flags(cls, expr):
256 flags = 0
257 if not expr:
258 return flags, expr
259 for idx, ch in enumerate(expr):
260 if ch not in cls.RE_FLAGS:
261 break
262 flags |= cls.RE_FLAGS[ch]
263 return flags, expr[idx:] if idx > 0 else expr
264
265 @classmethod
266 def _separate(cls, expr, delim=',', max_split=None, skip_delims=None):
267 if not expr:
268 return
269 counters = {k: 0 for k in _MATCHING_PARENS.values()}
270 start, splits, pos, delim_len = 0, 0, 0, len(delim) - 1
271 in_quote, escaping, skipping = None, False, 0
272 after_op, in_regex_char_group, skip_re = True, False, 0
273
274 for idx, char in enumerate(expr):
275 if skip_re > 0:
276 skip_re -= 1
277 continue
278 if not in_quote:
279 if char in _MATCHING_PARENS:
280 counters[_MATCHING_PARENS[char]] += 1
281 elif char in counters:
282 counters[char] -= 1
283 if not escaping and char in _QUOTES and in_quote in (char, None):
284 if in_quote or after_op or char != '/':
285 in_quote = None if in_quote and not in_regex_char_group else char
286 if in_quote is None and char == '/' and delim != '/':
287 # regexp flags
288 n_idx = idx + 1
289 while n_idx < len(expr) and expr[n_idx] in cls.RE_FLAGS:
290 n_idx += 1
291 skip_re = n_idx - idx - 1
292 if skip_re > 0:
293 continue
294 elif in_quote == '/' and char in '[]':
295 in_regex_char_group = char == '['
296 escaping = not escaping and in_quote and char == '\\'
297 after_op = not in_quote and char in cls.OP_CHARS or (char == ' ' and after_op)
298
299 if char != delim[pos] or any(counters.values()) or in_quote:
300 pos = skipping = 0
301 continue
302 elif skipping > 0:
303 skipping -= 1
304 continue
305 elif pos == 0 and skip_delims:
306 here = expr[idx:]
307 for s in skip_delims if isinstance(skip_delims, (list, tuple)) else [skip_delims]:
308 if here.startswith(s) and s:
309 skipping = len(s) - 1
310 break
311 if skipping > 0:
312 continue
313 if pos < delim_len:
314 pos += 1
315 continue
316 yield expr[start: idx - delim_len]
317 start, pos = idx + 1, 0
318 splits += 1
319 if max_split and splits >= max_split:
320 break
321 yield expr[start:]
322
323 @classmethod
324 def _separate_at_paren(cls, expr, delim):
325 separated = list(cls._separate(expr, delim, 1))
326
327 if len(separated) < 2:
328 raise cls.Exception('No terminating paren {delim} in {expr}'.format(**locals()))
329 return separated[0][1:].strip(), separated[1].strip()
330
331 @staticmethod
332 def _all_operators():
333 return itertools.chain(
334 # Ref: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Operator_Precedence
335 _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS)
336
337 def _operator(self, op, left_val, right_expr, expr, local_vars, allow_recursion):
338 if op in ('||', '&&'):
339 if (op == '&&') ^ _ternary(left_val):
340 return left_val # short circuiting
341 elif op == '??':
342 if left_val not in (None, self.undefined):
343 return left_val
344 elif op == '?':
345 right_expr = _ternary(left_val, *self._separate(right_expr, ':', 1))
346
347 right_val = self.interpret_expression(right_expr, local_vars, allow_recursion)
348 opfunc = op and next((v for k, v in self._all_operators() if k == op), None)
349 if not opfunc:
350 return right_val
351
352 try:
353 return opfunc(left_val, right_val)
354 except Exception as e:
355 raise self.Exception('Failed to evaluate {left_val!r} {op} {right_val!r}'.format(**locals()), expr, cause=e)
356
357 def _index(self, obj, idx, allow_undefined=False):
358 if idx == 'length':
359 return len(obj)
360 try:
361 return obj[int(idx)] if isinstance(obj, list) else obj[idx]
362 except Exception as e:
363 if allow_undefined:
364 return self.undefined
365 raise self.Exception('Cannot get index {idx}'.format(**locals()), expr=repr(obj), cause=e)
366
367 def _dump(self, obj, namespace):
368 try:
369 return json.dumps(obj)
370 except TypeError:
371 return self._named_object(namespace, obj)
372
373 def interpret_statement(self, stmt, local_vars, allow_recursion=100):
374 if allow_recursion < 0:
375 raise self.Exception('Recursion limit reached')
376 allow_recursion -= 1
377
378 should_return = False
379 sub_statements = list(self._separate(stmt, ';')) or ['']
380 expr = stmt = sub_statements.pop().strip()
381 for sub_stmt in sub_statements:
382 ret, should_return = self.interpret_statement(sub_stmt, local_vars, allow_recursion)
383 if should_return:
384 return ret, should_return
385
386 m = re.match(r'(?P<var>(?:var|const|let)\s)|return(?:\s+|(?=["\'])|$)|(?P<throw>throw\s+)', stmt)
387 if m:
388 expr = stmt[len(m.group(0)):].strip()
389 if m.group('throw'):
390 raise JS_Throw(self.interpret_expression(expr, local_vars, allow_recursion))
391 should_return = not m.group('var')
392 if not expr:
393 return None, should_return
394
395 if expr[0] in _QUOTES:
396 inner, outer = self._separate(expr, expr[0], 1)
397 if expr[0] == '/':
398 flags, _ = self._regex_flags(outer)
399 inner, outer = inner.replace('"', r'\"'), ''
400 inner = re.compile(js_to_json(inner + expr[0]), flags=flags) # , strict=True))
401 else:
402 inner = json.loads(js_to_json(inner + expr[0])) # , strict=True))
403 if not outer:
404 return inner, should_return
405 expr = self._named_object(local_vars, inner) + outer
406
407 if expr.startswith('new '):
408 obj = expr[4:]
409 if obj.startswith('Date('):
410 left, right = self._separate_at_paren(obj[4:], ')')
411 expr = unified_timestamp(
412 self.interpret_expression(left, local_vars, allow_recursion), False)
413 if not expr:
414 raise self.Exception('Failed to parse date {left!r}'.format(**locals()), expr=expr)
415 expr = self._dump(int(expr * 1000), local_vars) + right
416 else:
417 raise self.Exception('Unsupported object {obj}'.format(**locals()), expr=expr)
418
419 if expr.startswith('void '):
420 left = self.interpret_expression(expr[5:], local_vars, allow_recursion)
421 return None, should_return
422
423 if expr.startswith('{'):
424 inner, outer = self._separate_at_paren(expr, '}')
425 # try for object expression
426 sub_expressions = [list(self._separate(sub_expr.strip(), ':', 1)) for sub_expr in self._separate(inner)]
427 if all(len(sub_expr) == 2 for sub_expr in sub_expressions):
428 return dict(
429 (key_expr if re.match(_NAME_RE, key_expr) else key_expr,
430 self.interpret_expression(val_expr, local_vars, allow_recursion))
431 for key_expr, val_expr in sub_expressions), should_return
432 # or statement list
433 inner, should_abort = self.interpret_statement(inner, local_vars, allow_recursion)
434 if not outer or should_abort:
435 return inner, should_abort or should_return
436 else:
437 expr = self._dump(inner, local_vars) + outer
438
439 if expr.startswith('('):
440 inner, outer = self._separate_at_paren(expr, ')')
441 inner, should_abort = self.interpret_statement(inner, local_vars, allow_recursion)
442 if not outer or should_abort:
443 return inner, should_abort or should_return
444 else:
445 expr = self._dump(inner, local_vars) + outer
446
447 if expr.startswith('['):
448 inner, outer = self._separate_at_paren(expr, ']')
449 name = self._named_object(local_vars, [
450 self.interpret_expression(item, local_vars, allow_recursion)
451 for item in self._separate(inner)])
452 expr = name + outer
453
454 m = re.match(r'''(?x)
455 (?P<try>try|finally)\s*|
456 (?P<catch>catch\s*(?P<err>\(\s*{_NAME_RE}\s*\)))|
457 (?P<switch>switch)\s*\(|
458 (?P<for>for)\s*\(|'''.format(**globals()), expr)
459 md = m.groupdict() if m else {}
460 if md.get('try'):
461 if expr[m.end()] == '{':
462 try_expr, expr = self._separate_at_paren(expr[m.end():], '}')
463 else:
464 try_expr, expr = expr[m.end() - 1:], ''
465 try:
466 ret, should_abort = self.interpret_statement(try_expr, local_vars, allow_recursion)
467 if should_abort:
468 return ret, True
469 except JS_Throw as e:
470 local_vars[self._EXC_NAME] = e.error
471 except Exception as e:
472 # XXX: This works for now, but makes debugging future issues very hard
473 local_vars[self._EXC_NAME] = e
474 ret, should_abort = self.interpret_statement(expr, local_vars, allow_recursion)
475 return ret, should_abort or should_return
476
477 elif md.get('catch'):
478 catch_expr, expr = self._separate_at_paren(expr[m.end():], '}')
479 if self._EXC_NAME in local_vars:
480 catch_vars = local_vars.new_child({m.group('err'): local_vars.pop(self._EXC_NAME)})
481 ret, should_abort = self.interpret_statement(catch_expr, catch_vars, allow_recursion)
482 if should_abort:
483 return ret, True
484
485 ret, should_abort = self.interpret_statement(expr, local_vars, allow_recursion)
486 return ret, should_abort or should_return
487
488 elif md.get('for'):
489 constructor, remaining = self._separate_at_paren(expr[m.end() - 1:], ')')
490 if remaining.startswith('{'):
491 body, expr = self._separate_at_paren(remaining, '}')
492 else:
493 switch_m = re.match(r'switch\s*\(', remaining) # FIXME
494 if switch_m:
495 switch_val, remaining = self._separate_at_paren(remaining[switch_m.end() - 1:], ')')
496 body, expr = self._separate_at_paren(remaining, '}')
497 body = 'switch(%s){%s}' % (switch_val, body)
498 else:
499 body, expr = remaining, ''
500 start, cndn, increment = self._separate(constructor, ';')
501 self.interpret_expression(start, local_vars, allow_recursion)
502 while True:
503 if not _ternary(self.interpret_expression(cndn, local_vars, allow_recursion)):
504 break
505 try:
506 ret, should_abort = self.interpret_statement(body, local_vars, allow_recursion)
507 if should_abort:
508 return ret, True
509 except JS_Break:
510 break
511 except JS_Continue:
512 pass
513 self.interpret_expression(increment, local_vars, allow_recursion)
514 ret, should_abort = self.interpret_statement(expr, local_vars, allow_recursion)
515 return ret, should_abort or should_return
516
517 elif md.get('switch'):
518 switch_val, remaining = self._separate_at_paren(expr[m.end() - 1:], ')')
519 switch_val = self.interpret_expression(switch_val, local_vars, allow_recursion)
520 body, expr = self._separate_at_paren(remaining, '}')
521 items = body.replace('default:', 'case default:').split('case ')[1:]
522 for default in (False, True):
523 matched = False
524 for item in items:
525 case, stmt = (i.strip() for i in self._separate(item, ':', 1))
526 if default:
527 matched = matched or case == 'default'
528 elif not matched:
529 matched = (case != 'default'
530 and switch_val == self.interpret_expression(case, local_vars, allow_recursion))
531 if not matched:
532 continue
533 try:
534 ret, should_abort = self.interpret_statement(stmt, local_vars, allow_recursion)
535 if should_abort:
536 return ret
537 except JS_Break:
538 break
539 if matched:
540 break
541 ret, should_abort = self.interpret_statement(expr, local_vars, allow_recursion)
542 return ret, should_abort or should_return
543
544 # Comma separated statements
545 sub_expressions = list(self._separate(expr))
546 if len(sub_expressions) > 1:
547 for sub_expr in sub_expressions:
548 ret, should_abort = self.interpret_statement(sub_expr, local_vars, allow_recursion)
549 if should_abort:
550 return ret, True
551 return ret, False
552
553 for m in re.finditer(r'''(?x)
554 (?P<pre_sign>\+\+|--)(?P<var1>{_NAME_RE})|
555 (?P<var2>{_NAME_RE})(?P<post_sign>\+\+|--)'''.format(**globals()), expr):
556 var = m.group('var1') or m.group('var2')
557 start, end = m.span()
558 sign = m.group('pre_sign') or m.group('post_sign')
559 ret = local_vars[var]
560 local_vars[var] += 1 if sign[0] == '+' else -1
561 if m.group('pre_sign'):
562 ret = local_vars[var]
563 expr = expr[:start] + self._dump(ret, local_vars) + expr[end:]
564
565 if not expr:
566 return None, should_return
567
568 m = re.match(r'''(?x)
569 (?P<assign>
570 (?P<out>{_NAME_RE})(?:\[(?P<index>[^\]]+?)\])?\s*
571 (?P<op>{_OPERATOR_RE})?
572 =(?!=)(?P<expr>.*)$
573 )|(?P<return>
574 (?!if|return|true|false|null|undefined)(?P<name>{_NAME_RE})$
575 )|(?P<indexing>
576 (?P<in>{_NAME_RE})\[(?P<idx>.+)\]$
577 )|(?P<attribute>
578 (?P<var>{_NAME_RE})(?:(?P<nullish>\?)?\.(?P<member>[^(]+)|\[(?P<member2>[^\]]+)\])\s*
579 )|(?P<function>
580 (?P<fname>{_NAME_RE})\((?P<args>.*)\)$
581 )'''.format(**globals()), expr)
582 md = m.groupdict() if m else {}
583 if md.get('assign'):
584 left_val = local_vars.get(m.group('out'))
585
586 if not m.group('index'):
587 local_vars[m.group('out')] = self._operator(
588 m.group('op'), left_val, m.group('expr'), expr, local_vars, allow_recursion)
589 return local_vars[m.group('out')], should_return
590 elif left_val in (None, self.undefined):
591 raise self.Exception('Cannot index undefined variable ' + m.group('out'), expr=expr)
592
593 idx = self.interpret_expression(m.group('index'), local_vars, allow_recursion)
594 if not isinstance(idx, (int, float)):
595 raise self.Exception('List index %s must be integer' % (idx, ), expr=expr)
596 idx = int(idx)
597 left_val[idx] = self._operator(
598 m.group('op'), self._index(left_val, idx), m.group('expr'), expr, local_vars, allow_recursion)
599 return left_val[idx], should_return
600
601 elif expr.isdigit():
602 return int(expr), should_return
603
604 elif expr == 'break':
605 raise JS_Break()
606 elif expr == 'continue':
607 raise JS_Continue()
608
609 elif expr == 'undefined':
610 return self.undefined, should_return
611
612 elif md.get('return'):
613 return local_vars[m.group('name')], should_return
614
615 try:
616 ret = json.loads(js_to_json(expr)) # strict=True)
617 if not md.get('attribute'):
618 return ret, should_return
619 except ValueError:
620 pass
621
622 if md.get('indexing'):
623 val = local_vars[m.group('in')]
624 idx = self.interpret_expression(m.group('idx'), local_vars, allow_recursion)
625 return self._index(val, idx), should_return
626
627 for op, _ in self._all_operators():
628 # hackety: </> have higher priority than <</>>, but don't confuse them
629 skip_delim = (op + op) if op in '<>*?' else None
630 if op == '?':
631 skip_delim = (skip_delim, '?.')
632 separated = list(self._separate(expr, op, skip_delims=skip_delim))
633 if len(separated) < 2:
634 continue
635
636 right_expr = separated.pop()
637 while op == '-' and len(separated) > 1 and not separated[-1].strip():
638 right_expr = '-' + right_expr
639 separated.pop()
640 left_val = self.interpret_expression(op.join(separated), local_vars, allow_recursion)
641 return self._operator(op, left_val, right_expr, expr, local_vars, allow_recursion), should_return
642
643 if md.get('attribute'):
644 variable, member, nullish = m.group('var', 'member', 'nullish')
645 if not member:
646 member = self.interpret_expression(m.group('member2'), local_vars, allow_recursion)
647 arg_str = expr[m.end():]
648 if arg_str.startswith('('):
649 arg_str, remaining = self._separate_at_paren(arg_str, ')')
650 else:
651 arg_str, remaining = None, arg_str
652
653 def assertion(cndn, msg):
654 """ assert, but without risk of getting optimized out """
655 if not cndn:
656 raise ExtractorError('{member} {msg}'.format(**locals()), expr=expr)
657
658 def eval_method():
659 if (variable, member) == ('console', 'debug'):
660 return
661 types = {
662 'String': compat_str,
663 'Math': float,
664 }
665 obj = local_vars.get(variable)
666 if obj in (self.undefined, None):
667 obj = types.get(variable, self.undefined)
668 if obj is self.undefined:
669 try:
670 if variable not in self._objects:
671 self._objects[variable] = self.extract_object(variable)
672 obj = self._objects[variable]
673 except self.Exception:
674 if not nullish:
675 raise
676
677 if nullish and obj is self.undefined:
678 return self.undefined
679
680 # Member access
681 if arg_str is None:
682 return self._index(obj, member, nullish)
683
684 # Function call
685 argvals = [
686 self.interpret_expression(v, local_vars, allow_recursion)
687 for v in self._separate(arg_str)]
688
689 if obj == compat_str:
690 if member == 'fromCharCode':
691 assertion(argvals, 'takes one or more arguments')
692 return ''.join(map(chr, argvals))
693 raise self.Exception('Unsupported string method ' + member, expr=expr)
694 elif obj == float:
695 if member == 'pow':
696 assertion(len(argvals) == 2, 'takes two arguments')
697 return argvals[0] ** argvals[1]
698 raise self.Exception('Unsupported Math method ' + member, expr=expr)
699
700 if member == 'split':
701 assertion(argvals, 'takes one or more arguments')
702 assertion(len(argvals) == 1, 'with limit argument is not implemented')
703 return obj.split(argvals[0]) if argvals[0] else list(obj)
704 elif member == 'join':
705 assertion(isinstance(obj, list), 'must be applied on a list')
706 assertion(len(argvals) == 1, 'takes exactly one argument')
707 return argvals[0].join(obj)
708 elif member == 'reverse':
709 assertion(not argvals, 'does not take any arguments')
710 obj.reverse()
711 return obj
712 elif member == 'slice':
713 assertion(isinstance(obj, list), 'must be applied on a list')
714 assertion(len(argvals) == 1, 'takes exactly one argument')
715 return obj[argvals[0]:]
716 elif member == 'splice':
717 assertion(isinstance(obj, list), 'must be applied on a list')
718 assertion(argvals, 'takes one or more arguments')
719 index, howMany = map(int, (argvals + [len(obj)])[:2])
720 if index < 0:
721 index += len(obj)
722 add_items = argvals[2:]
723 res = []
724 for i in range(index, min(index + howMany, len(obj))):
725 res.append(obj.pop(index))
726 for i, item in enumerate(add_items):
727 obj.insert(index + i, item)
728 return res
729 elif member == 'unshift':
730 assertion(isinstance(obj, list), 'must be applied on a list')
731 assertion(argvals, 'takes one or more arguments')
732 for item in reversed(argvals):
733 obj.insert(0, item)
734 return obj
735 elif member == 'pop':
736 assertion(isinstance(obj, list), 'must be applied on a list')
737 assertion(not argvals, 'does not take any arguments')
738 if not obj:
739 return
740 return obj.pop()
741 elif member == 'push':
742 assertion(argvals, 'takes one or more arguments')
743 obj.extend(argvals)
744 return obj
745 elif member == 'forEach':
746 assertion(argvals, 'takes one or more arguments')
747 assertion(len(argvals) <= 2, 'takes at-most 2 arguments')
748 f, this = (argvals + [''])[:2]
749 return [f((item, idx, obj), {'this': this}, allow_recursion) for idx, item in enumerate(obj)]
750 elif member == 'indexOf':
751 assertion(argvals, 'takes one or more arguments')
752 assertion(len(argvals) <= 2, 'takes at-most 2 arguments')
753 idx, start = (argvals + [0])[:2]
754 try:
755 return obj.index(idx, start)
756 except ValueError:
757 return -1
758
759 idx = int(member) if isinstance(obj, list) else member
760 return obj[idx](argvals, allow_recursion=allow_recursion)
761
762 if remaining:
763 ret, should_abort = self.interpret_statement(
764 self._named_object(local_vars, eval_method()) + remaining,
765 local_vars, allow_recursion)
766 return ret, should_return or should_abort
767 else:
768 return eval_method(), should_return
769
770 elif md.get('function'):
771 fname = m.group('fname')
772 argvals = [self.interpret_expression(v, local_vars, allow_recursion)
773 for v in self._separate(m.group('args'))]
774 if fname in local_vars:
775 return local_vars[fname](argvals, allow_recursion=allow_recursion), should_return
776 elif fname not in self._functions:
777 self._functions[fname] = self.extract_function(fname)
778 return self._functions[fname](argvals, allow_recursion=allow_recursion), should_return
779
780 raise self.Exception(
781 'Unsupported JS expression ' + (expr[:40] if expr != stmt else ''), expr=stmt)
782
783 def interpret_expression(self, expr, local_vars, allow_recursion):
784 ret, should_return = self.interpret_statement(expr, local_vars, allow_recursion)
785 if should_return:
786 raise self.Exception('Cannot return from an expression', expr)
787 return ret
788
789 def extract_object(self, objname):
790 _FUNC_NAME_RE = r'''(?:[a-zA-Z$0-9]+|"[a-zA-Z$0-9]+"|'[a-zA-Z$0-9]+')'''
791 obj = {}
792 obj_m = re.search(
793 r'''(?x)
794 (?<!this\.)%s\s*=\s*{\s*
795 (?P<fields>(%s\s*:\s*function\s*\(.*?\)\s*{.*?}(?:,\s*)?)*)
796 }\s*;
797 ''' % (re.escape(objname), _FUNC_NAME_RE),
798 self.code)
799 if not obj_m:
800 raise self.Exception('Could not find object ' + objname)
801 fields = obj_m.group('fields')
802 # Currently, it only supports function definitions
803 fields_m = re.finditer(
804 r'''(?x)
805 (?P<key>%s)\s*:\s*function\s*\((?P<args>(?:%s|,)*)\){(?P<code>[^}]+)}
806 ''' % (_FUNC_NAME_RE, _NAME_RE),
807 fields)
808 for f in fields_m:
809 argnames = self.build_arglist(f.group('args'))
810 obj[remove_quotes(f.group('key'))] = self.build_function(argnames, f.group('code'))
811
812 return obj
813
814 def extract_function_code(self, funcname):
815 """ @returns argnames, code """
816 func_m = re.search(
817 r'''(?xs)
818 (?:
819 function\s+%(name)s|
820 [{;,]\s*%(name)s\s*=\s*function|
821 (?:var|const|let)\s+%(name)s\s*=\s*function
822 )\s*
823 \((?P<args>[^)]*)\)\s*
824 (?P<code>{.+})''' % {'name': re.escape(funcname)},
825 self.code)
826 code, _ = self._separate_at_paren(func_m.group('code'), '}') # refine the match
827 if func_m is None:
828 raise self.Exception('Could not find JS function "{funcname}"'.format(**locals()))
829 return self.build_arglist(func_m.group('args')), code
830
831 def extract_function(self, funcname):
832 return self.extract_function_from_code(*self.extract_function_code(funcname))
833
834 def extract_function_from_code(self, argnames, code, *global_stack):
835 local_vars = {}
836 while True:
837 mobj = re.search(r'function\((?P<args>[^)]*)\)\s*{', code)
838 if mobj is None:
839 break
840 start, body_start = mobj.span()
841 body, remaining = self._separate_at_paren(code[body_start - 1:], '}')
842 name = self._named_object(
843 local_vars,
844 self.extract_function_from_code(
845 self.build_arglist(mobj.group('args')),
846 body, local_vars, *global_stack))
847 code = code[:start] + name + remaining
848 return self.build_function(argnames, code, local_vars, *global_stack)
849
850 def call_function(self, funcname, *args):
851 return self.extract_function(funcname)(args)
852
853 @classmethod
854 def build_arglist(cls, arg_text):
855 if not arg_text:
856 return []
857
858 def valid_arg(y):
859 y = y.strip()
860 if not y:
861 raise cls.Exception('Missing arg in "%s"' % (arg_text, ))
862 return y
863
864 return [valid_arg(x) for x in cls._separate(arg_text)]
865
866 def build_function(self, argnames, code, *global_stack):
867 global_stack = list(global_stack) or [{}]
868 argnames = tuple(argnames)
869
870 def resf(args, kwargs={}, allow_recursion=100):
871 global_stack[0].update(
872 zip_longest(argnames, args, fillvalue=None))
873 global_stack[0].update(kwargs)
874 var_stack = LocalNameSpace(*global_stack)
875 ret, should_abort = self.interpret_statement(code.replace('\n', ''), var_stack, allow_recursion - 1)
876 if should_abort:
877 return ret
878 return resf
|